LabSENSE

A laboratory IoT monitoring and analytics platform for chemistry lab environments. Collects real-time sensor data from Raspberry Pi devices, retrieves chemical inventory data from a cloud API, processes hazardous waste records, and generates interactive HTML dashboards visualising all data streams.

Architecture

[Raspberry Pi 4 / Pico]  →  MQTT  →  Subscriber  →  SQL Server (labsense DB)
[ChemInventory API]      →  HTTP  →  CI scripts   →  SQL Server
[Waste Master.xlsx]      →  file  →  Waste module →  HTML dashboards (plots/)
[Balance (RS-232)]       →  serial → balance_ci   →  ChemInventory API update

Modules

Directory	Purpose
Labsense_SQL/	SQL Server data layer — inserts, queries, and HTML dashboard generation for chemicals, electricity, fumehood, and water
Labsense_Sensors/	Raspberry Pi sensor scripts — VL53L1X ToF (fumehood sash position), LTR559 light sensor, and GPIO water flow sensors
ChemInventory/	Ad-hoc scripts querying the ChemInventory REST API by CAS number or GHS hazard code
Waste/	Reads Waste Master.xlsx, proportionally allocates waste volumes across HP1–HP15 hazard codes, and generates per-HP dashboards
Balance Comms/	Scans a barcode, reads weight from a serial balance (Denver Instruments SI-2002), and updates ChemInventory via REST API
RaspberryPi-4/	Scheduled scripts pushing sensor streams (electricity, fumehood, GHS, order/waste lists) to Azure IoT Hub
RaspberryPi-Pico/	MicroPython firmware — WiFi, Azure IoT Hub connection, BME68x environmental sensor
Labsense_Excel/	Legacy Excel-based order and waste update scripts using openpyxl
tests/	pytest unit tests for core processing logic
plots/	Generated HTML dashboards and matplotlib PNGs
create_main_dashboard.py	Generates the top-level HTML landing page linking all dashboards in plots/

Dependencies

Managed via Conda. Key packages:

• Data processing: pandas, numpy, matplotlib
• Database: pyodbc (SQL Server, ODBC Driver 18)
• Messaging: paho-mqtt, Azure IoT Hub SDK
• APIs: requests
• Spreadsheets: openpyxl
• Configuration: python-dotenv
• Testing: pytest
• Hardware (Raspberry Pi only): RPi.GPIO, gpiozero, VL53L1X, ltr559, hx711-multi, adafruit-circuitpython-charlcd

See environment.yml for the full list.

Installation

git clone https://github.com/yourusername/labsense.git
cd labsense
conda env create -f environment.yml
conda activate labsense

Configuration

All SQL/processing scripts read configuration from Labsense_SQL/.env:

# EmonCMS API key (electricity/water consumption data)
EMONCMS_API_KEY=your_emoncms_api_key

# EmonCMS base URL (host/IP + scheme)
EMONCMS_BASE_URL=https://your_emoncms_host

# Logging level
LOG_LEVEL=INFO

# ChemInventory API token
CHEMINVENTORY_CONNECTION_STRING=your_cheminventory_api_token

# Toggle SQL Server inserts for ChemInventory data (True/False)
CHEMINVENTORY_INSERT_TO_SQL=True

# MQTT broker address (used by subscriber_sqlserver.py)
MQTT_SERVER=your_mqtt_broker_ip

# SQL Server connection
SQL_SERVER=your_sql_server_instance
SQL_DATABASE=labsense
SQL_TRUSTED_CONNECTION=yes
SQL_ENCRYPTION=Optional

Sensor scripts on the Raspberry Pi use Labsense_Sensors/.env for I2C addresses, sensor thresholds, and retry settings.

Running the Dashboards

Generate all HTML dashboards and the landing page:

python Labsense_SQL/ChemInventory_dashboard.py
python Labsense_SQL/consumption_dashboard.py
python Labsense_SQL/Fumehood_dashboard.py
python Labsense_SQL/water_dashboard.py
python Waste/processWasteMaster.py
python create_main_dashboard.py

Output is written to plots/. Open plots/summary_dashboard.html in a browser.

Running Tests

pytest